Getting Started

Based on the etcd v3.6 quickstart [5] [6] [7]. Commands assume Linux on amd64 and a shell.

Prerequisites

A Linux amd64 host with wget and tar.
Ports 2379 (client) and 2380 (peer) free on localhost.

Install

Download a release tarball and put the binaries on your path.

bash

ETCD_VER=v3.6.0
wget https://github.com/etcd-io/etcd/releases/download/${ETCD_VER}/etcd-${ETCD_VER}-linux-amd64.tar.gz
tar xzf etcd-${ETCD_VER}-linux-amd64.tar.gz
sudo mv etcd-${ETCD_VER}-linux-amd64/etcd* /usr/local/bin/
etcd --version && etcdctl version

To run it as a container instead, the official image is gcr.io/etcd-development/etcd (primary), with quay.io/coreos/etcd as a secondary mirror [7].

A first working setup

This runs a single-member etcd and stores a key.

Start the server. With no flags it listens for clients on 2379 and peers on 2380.
bash
```
etcd &
```
Write a key with etcdctl.
bash
```
etcdctl put greeting "Hello, etcd"
```

Expected output:

text

OK

Read it back.
bash
```
etcdctl get greeting
```

Expected output:

text

greeting
Hello, etcd

Verify it works

Check member health with the endpoint status command.

bash

etcdctl endpoint health

A healthy member reports that the endpoint is healthy along with a response time.

Where to go next

The single-member setup above is for local use only. For production concerns such as running a multi-member cluster, enabling TLS and RBAC, setting the storage quota, and backup and compaction, see the official documentation at etcd.io/docs. Do not run a single member as a real datastore; a cluster needs an odd number of members to tolerate failures.

Getting Started ​

Prerequisites ​

Install ​

A first working setup ​

Verify it works ​